Skip to content
IEEE TPAMI Trustworthy AI Special Issue

Submissions are now open for this IEEE TPAMI Special Issue.

Submissions open

Trustworthy AI in Pattern Recognition and Computer Vision: Attacks, Robustness, Safety, and Beyond

A TPAMI special issue on the safety, security, robustness, privacy and reliability of modern visual intelligence — from foundation models and multimodal LLMs to generative and embodied AI.

A Special Issue of IEEE Transactions on Pattern Analysis and Machine Intelligence

Submission deadline:

Submissions are handled on a rolling basis.

Submit Manuscript View Topics

The special issue

Aims & Scope

AI systems are now being deployed in safety-critical settings such as autonomous driving, medical diagnostics, biometric identification, content moderation, and AI-powered agents. Accuracy alone is no longer sufficient in these contexts. Systems must also be demonstrably trustworthy: robust to adversarial manipulation, protective of private data, fair across demographic groups, and safe under distribution shifts and unexpected inputs.

This special issue solicits original research, extended conference papers, and survey articles on the theory, methodology, and practice of trustworthy AI within pattern recognition and computer vision. Of particular interest is work that connects theoretical guarantees with practical deployment, that addresses the security challenges raised by large language and multimodal models, or that establishes rigorous evaluation frameworks for trustworthy AI systems.

We take a full-lifecycle view of visual AI safety, from data collection and training to evaluation, deployment, and interaction with users and environments.

Research areas

Topics of Interest

Submissions should make clear technical contributions to visual intelligence, including visual foundation models, multimodal models, generative visual models, and embodied AI.

Topics include, but are not limited to, the following areas.

Attacks and Threat Modelling

  • Adversarial examples: white-box, black-box, and physical-world attacks
  • Backdoor / trojan attacks on visual and multimodal models
  • Data poisoning, model supply-chain, and training-time attacks
  • Privacy attacks: membership inference, model inversion, attribute inference
  • Prompt injection and jailbreaking of LLMs and vision–language models
  • Visual prompt injection targeting multimodal AI agents
  • Deepfake generation and detection

Robustness and Defences

  • Adversarial training: AT, TRADES, and variants for vision models
  • Certified defences and formal verification of neural networks
  • Test-time defences, purification, and anomaly/OOD detection
  • Robustness of foundation models and zero/few-shot learners
  • Backdoor detection, Neural Cleanse, and model unlearning

Privacy-Preserving Learning

  • Federated learning: security, poisoning, and Byzantine robustness
  • Differential privacy in vision with utility–privacy tradeoff analysis
  • Unlearnable examples and machine unlearning
  • IP protection via adversarial watermarking and model fingerprinting

Fairness, Accountability, and Explainability

  • Bias detection and mitigation in visual recognition and generative AI
  • Interpretability and explainability for safety-critical vision systems
  • Auditing frameworks and third-party certification of AI models

AI Agent and LLM Safety

  • Security of autonomous AI agents: tool-call manipulation, environment injection
  • Hallucination detection and factual grounding of MLLMs
  • Red-teaming and safety benchmarks for foundation and agentic models
  • Safe reinforcement learning for embodied AI

Applications in High-Stakes Domains

  • Trustworthy biometrics under adversarial conditions
  • Robust medical image analysis
  • Security of perception systems in autonomous vehicles and robotics
  • Synthetic media detection for forensics and disinformation defence

Schedule

Important Dates

Submissions are handled on a rolling basis.

Manuscript submission deadline

Submit through the IEEE Submission Portal for TPAMI.

First-round review notification Target
Revision submission deadline Target
Final decision notification Target
Publication Target
Q1–Q2 2028

Review and publication dates are target milestones. Authors should follow the dates provided in their individual editorial decision letters.

For authors

Submission Instructions

Manuscripts should be submitted through the IEEE Submission Portal for TPAMI. All submissions must follow the TPAMI author guidelines and the journal’s standard review process.

  1. 1

    Prepare your manuscript

    Prepare your manuscript and supplementary materials according to the TPAMI author guidelines. Original research articles, extended conference papers, and high-quality survey articles within the scope of the special issue are welcome.

    TPAMI author guidelines
  2. 2

    Open the IEEE Submission Portal

    Open the IEEE Submission Portal for TPAMI and follow the on-screen instructions to prepare your submission.

    IEEE Submission Portal
  3. 3

    Select the special issue under Article Type

    In the Article Type field, select the following special issue:

    SI: Trustworthy AI in Pattern Recognition and Computer Vision: Attacks, Robustness, Safety, and Beyond
  • All submissions follow TPAMI’s standard editorial and peer-review procedures.

Submissions open

Submission deadline

1 March 2027

Submissions are handled on a rolling basis.

Submit Manuscript

IEEE Submission Portal

Extended conference papers

Extended conference papers are welcome provided that the submission contains at least 30% new content beyond the published conference version and explicitly references the prior version. Authors should also follow TPAMI’s guidance on substantial revisions of prior conference papers.

TPAMI Author Guidelines

Peer-review preference

The submission system offers single-anonymized and double-anonymized peer-review options. Please follow the on-screen instructions and prepare identifying information according to your selected option. The system states that this selection cannot be changed after submission.

Editorial team

Guest Editors

The guest editorial team brings together expertise in computer vision, pattern recognition, adversarial machine learning, and AI safety and security.

Portrait of Rama Chellappa

Lead Guest Editor

Rama Chellappa

Johns Hopkins University

Homepage
Portrait of Chun Pong Lau

Guest Editor

Chun Pong Lau

City University of Hong Kong

Homepage
Portrait of Zhongliang Guo

Guest Editor

Zhongliang Guo

University of Aberdeen

Homepage
Portrait of Jiang Liu

Guest Editor

Jiang Liu

Advanced Micro Devices, Inc. (AMD)

Homepage
Portrait of Linlin Shen

Guest Editor

Linlin Shen

Shenzhen University

Homepage
Portrait of Zhiyuan Zha

Guest Editor

Zhiyuan Zha

Jilin University

Homepage
Portrait of Yiming Li

Guest Editor

Yiming Li

Nanyang Technological University

Homepage

Contact

Contact the Guest Editorial Team

For questions about the scope of the special issue or general submission enquiries, please contact the Guest Editorial Team.

Guest Editorial Team

tpami.trustworthyai@gmail.com

For technical issues with the submission system, please use the support options available in the IEEE Submission Portal.

TPAMI Author Guidelines IEEE Submission Portal
Help us spread the word

Share this call

The special issue depends on reaching the right authors. If you work in adversarial ML, vision security, or LLM safety, please pass it on.

Share on X Share on LinkedIn

Suggested text

Call for Papers — IEEE TPAMI Special Issue on “Trustworthy AI in Pattern Recognition and Computer Vision: Attacks, Robustness, Safety, and Beyond”. Submissions are open and handled on a rolling basis. Submission deadline: 1 March 2027.